• Nenhum resultado encontrado

The contribution of GLONASS measurements to regional and continental scale geodetic monitoring regimes

N/A
N/A
Protected

Academic year: 2017

Share "The contribution of GLONASS measurements to regional and continental scale geodetic monitoring regimes"

Copied!
4
0
0

Texto

(1)

LETTER Earth Planets Space,52, 877–880, 2000

The contribution of GLONASS measurements to regional and continental scale

geodetic monitoring regimes

M. P. Stewart1, M. Tsakiri1, J. Wang1, and J. F. Monico2

1Curtin University of Technology, Western Australia 2FCT/UNESP, Brazil

(Received December 6, 1999; Revised May 12, 2000; Accepted May 15, 2000)

Throughout late 1998 and early 1999, the International GLONASS Experiment (IGEX) has delivered the first comprehensive inter-continental dual frequency GLONASS data set. This experiment represents the first opportunity to assess how a second global satellite positioning system could complement existing GPS geodetic infrastructure. Based on analysis of a three station network of IGEX stations from Southern Hemisphere IGEX stations, this paper discusses the internal and external precision of long baseline GPS, GLONASS and combined GPS/GLONASS solutions, and the possible contribution of GLONASS to future regional-scale geodetic work.

1.

Introduction

The tracking of GPS and GLONASS satellites as part of the International GLONASS Experiment (IGEX) officially commenced on the 20th October. For the duration of the experiment, over 60 GPS/GLONASS receivers have been located at sites around the world. Run under the auspices of the International Association of Geodesy (IAG), the Interna-tional GPS Service (IGS), the InternaInterna-tional Earth Rotation Service (IERS) and the Institute of Navigation (ION), this experiment represents the first opportunity to assess how a second global satellite positioning system could complement existing GPS infrastructure for geodetic purposes (Willis and Slater, 1999).

At the time of writing (November 1999), the GLONASS constellation has been reduced to 10 operating satellites, less than half the full constellation of 24. With the current eco-nomic situation in the Russian Federation and the aged status of the remaining constellation, the future of the GLONASS system is looking increasingly uncertain. The launch of three new GLONASS satellites in December 1998, and their sub-sequent coming on line in January/February 1999, led to a peak in the number of available satellites in the early months of the year. If the status of the GLONASS system continues to decline and with no alternative global positioning systems becoming available in the near future, the data collected dur-ing the IGEX campaign in the early months of 1999 will represent a unique data set for several years to come. Over this period, a sufficient number of GLONASS satellites (gen-erally greater than half the constellation) were operational to allow studies of the likely contribution secondary satellite po-sitioning systems could make to well established GPS-based global and regional monitoring regimes.

The availability of a second GPS-like global positioning

Copy right c The Society of Geomagnetism and Earth, Planetary and Space Sciences (SGEPSS); The Seismological Society of Japan; The Volcanological Society of Japan; The Geodetic Society of Japan; The Japanese Society for Planetary Sciences.

system would have two significant benefits to geodetic mon-itoring. First, solutions from a second system could provide an independent check on GPS solutions, thus improving qual-ity control. Second, raw observations from the second sys-tem could be combined directly with raw GPS observations to provide a geometrically stronger solution (GPS is known to be geometrically weaker in height and at high latitudes).

In this paper, we illustrate results from a subset of the southern hemisphere network of IGEX stations described in Stewart et al.(1999), in the context of the comparison of stand-alone GPS and GLONASS solutions, and solutions from direct combination of raw GPS/GLONASS observa-tions.

2.

Regional Observation Network

The sub-network comprises three stations, Yarragadee and Mt. Stromlo, both located in Australia, and Crary Science Laboratory, McMurdo, Antarctica. The Australian stations used Ashtech Z18 dual frequency receivers whilst a Javad Positioning Systems Legacy dual frequency receiver was in-stalled at Crary. Figure 1 shows the regional sub-network de-scribed in this paper. Baseline lengths range from 3,199 km (Yarragadee–Mt. Stromlo) to 5,558 km (Yarragadee–Crary). Each of the three baselines was processed using a sep-arate week of data. Allocation of one baseline per week ensured that separate baseline solutions were independent. Therefore, the baseline network could be subsequently sub-jected to a least squares network adjustment without consid-ering baseline correlations. Solutions were computed on a daily basis and then combined into multi-arc solutions using the CODE precise GPS and GLONASS orbits com-puted by the Centre for Orbit Determination (CODE), Berne (Ineichen et al., 1999). GPS-only, GLONASS-only and combined GPS/GLONASS solutions were computed. Due to limitations on the length of this paper, the reader is re-ferred to Stewart et al.(1999) for full details of the basic processing methodology for the GPS-only and

(2)

878 M. P. STEWARTet al.: REGIONAL AND CONTINENTAL GLONASS MONITORING

Fig. 1. Southern hemisphere region sub-network.

only solutions.

3.

Combined GPS/GLONASS Solutions

Raw GPS and GLONASS carrier phase observations are fundamentally similar, with the main difference being that whilst GPS satellites transmit on the same L-band frequen-cies, GLONASS satellites transmit on different, albeit sim-ilar, frequencies. Therefore, the double differenced observ-able formed from raw GLONASS observations still contains some clock biases (e.g. Raby and Daly, 1994). One solution to this problem is to scale the GLONASS L1 and L2 ob-servations from each satellite to a common frequency (e.g. Leick, 1998). In the data processing presented in this paper, GLONASS L1 and L2 carrier phase observations were scaled to the GPS L1 and L2 frequencies respectively.

Once scaled to the GPS L1 and L2 frequencies, GLONASS carrier phase observations, in terms of data processing, may be treated as if they were GPS observations, with three major qualifiers. First, the integer nature of the GLONASS double difference ambiguity is lost. However, as the ionospheric free linear combination of L1 and L2 for GPS also results in a situation where ambiguities and cycle slips are non-integer in nature, on long baselines this factor is not an issue. Second, a difference exists between the reference frame in which the GLONASS satellite orbits are computed and the GPS satellite orbits are computed. Finally, a time offset between GPS and GLONASS time systems must be accounted for.

Point three above has by far the most critical effect when directly combining GPS and GLONASS observations. With-out the availability of highly accurate values for the time offset between the two systems, it is necessary to sacrifice an observation to solve for this offset. In the processing described here, only GPS–GPS and GLONASS–GLONASS double difference observations were formed in the combined solution, thus negating between system time biases.

4.

Results

Table 1 gives the individual statistics for the three multi-arc baseline solutions. It can be seen that the number of double difference satellite observations increases by between 20%

and 40% when GLONASS is included in the full solution. The estimated noise on the double difference observations is increased marginally for the combined solution, though not sufficiently to indicate that GLONASS observations are sig-nificantly noisier than GPS observations. The associated pre-cision of the individual baseline components (σxyz) generally decreases, indicating improved geometric strength in the solution when the additional GLONASS observations are included. Note that these statistics are only internal pre-cision indicators i.e. they are derived from the final solution covariance matrix. It is generally accepted that GPS covari-ance matrices tend to derive error estimates that are too op-timistic and the error estimates for the GPS, GLONASS and combined GPS/GLONASS solutions presented here are no exception. These precisions are presented rather to demon-state the improvement in geometric strength of the com-bined GPS/GLONASS solutions over the individual GPS and GLONASS solutions. Baseline lengths are in agreement for all three type of solution at the 0.01ppm level, demonstrating that even with an incomplete constellation, GLONASS-only solutions are comparable to GPS. The higher uncertainty on the baseline components from the GLONASS-only solutions represents the lower number of double difference GLONASS observations that could be formed in comparison to GPS.

It is also evident from Table 1 that the internal precision of the baseline from Mt. Stromlo to Crary is worse than the other two lines. This is attributed to a lower number of GPS observations available over the processed multi-day arc for that baseline. In GPS week 0992, the Z18 receiver at Stromlo recorded only between 12 and 16 hours of data per day. The outage did not reduce the number of GLONASS observations available, however (in fact, more GLONASS measurements were observed in total on this line than Yarragadee–Crary). It is also possible that this additional noise is caused by more inclement antarctic atmospheric conditions in week 992, al-though further analysis would be required to justify this state-ment. Baseline Yarragadee to Stromlo was processed for week 991 during which no such outage problems occurred.

(3)

or-M. P. STEWARTet al.: REGIONAL AND CONTINENTAL GLONASS MONITORING 879

Table 1. Baseline statistics from GPS, GLONASS and GPS/GLONASS processing.

Yarragadee–Stromlo

solution type No. obs. σx(m) σy(m) σz(m) obs. noise Length (m)

GPS 6391 0.009 0.010 0.007 4.3 mm 3199303.737

GLONASS 2005 0.015 0.025 0.016 4.6 mm 3199303.736

GPS+GLONASS 8396 0.005 0.007 0.005 4.7 mm 3199303.695

Yarragadee–Crary

solution type No. obs. σx(m) σy(m) σz(m) obs. noise Length (m)

GPS 5896 0.006 0.010 0.011 5.5 mm 5778302.223

GLONASS 1034 0.020 0.039 0.042 5.5 mm 5778302.177

GPS+GLONASS 6930 0.004 0.009 0.010 6.4 mm 5778302.204

Mt. Stromlo–Crary

solution type No. obs. σx(m) σy(m) σz(m) obs. noise Length (m)

GPS 3628 0.014 0.007 0.018 6.7 mm 4698509.709

GLONASS 1783 0.035 0.023 0.029 5.0 mm 4698509.730

GPS+GLONASS 5411 0.023 0.011 0.027 7.0 mm 4698509.700

Table 2. Standard deviations from least squares network adjustment from GPS-only, GLONASS-only and GPS/GLONASS solutions.

Station σx(m) σy(m) σz(m) σx(m) σy(m) σz(m) σx(m) σy(m) σz(m)

GPS GLONASS GPS/GLONASS

Crary 0.015 0.020 0.031 0.021 0.028 0.028 0.036 0.046 0.061

Mt. Stromlo 0.002 0.013 0.022 0.015 0.021 0.015 0.028 0.038 0.026

bits used for the analysis presented here are restricted to a smaller number of tracking stations than in the northern hemi-sphere, reference frame definition for the IGEX GLONASS orbits may be more uncertain in the Southern Hemisphere.

5.

Concluding Remarks

The main difficulty with processing data from the IGEX experiment on baselines of several thousand kilometres in length is the lack of available GLONASS satellites with which to form double difference observations. Taking into account that only fixed precise ephemerides and basic mod-elling techniques were applied, results from the adjusted GLONASS-only solution are encouraging. It seems likely that the larger error ellipses in the GLONASS least squares solution are due to the availability of fewer GLONASS ob-servations. GLONASS solutions demonstrate similar char-acteristics to GPS and, if fully operational, could be used as an independent alternative to GPS for geodetic purposes.

Even with a restricted GLONASS constellation, baseline lengths from the two positioning systems have been shown to agree at the 0.01 ppm level. Although other space geodetic techniques can be used for independent validation of regional GPS networks, in practical terms, only GLONASS (or sim-ilar future systems) could provide additional information at

each monitoring points. GPS and GLONASS solutions may never be completely independent (for example, in the IGEX campaign observations were taken by the same receiver to the same antenna) but regional solutions from two separate systems could provide an indication of systematic biases, in a field where very small changes in station coordinates are often interpreted as tectonic motion.

(4)

880 M. P. STEWARTet al.: REGIONAL AND CONTINENTAL GLONASS MONITORING

the existing IGEX network, some form of downweighting is necessary for the GLONASS observations to account for GLONASS orbital uncertainties. Conversely, it may be an-ticipated that this problem could be eliminated if (or when) GLONASS precise orbital computation approaches a similar level of resolution as GPS precise orbits.

Acknowledgments. The authors would like to thank the reviewers for some constructive comments regarding an earlier draft of this paper.

References

Ineichen, D., M. Rothatcher, T. Springer, and G. Beutler, Results of CODE as an Analysis Centre of the IGEX-98 Campaign, IGEX-99 Workshop, September 13–14, Nashville, Tennessee, U.S.A., 1999.

Leick, A., GLONASS satellite surveying,Journal of Surveying Engineering, 121, 91–99, 1998.

Raby, P. and P. Daly, Surveying with GLONASS: Calibration, Error Sources and Results,Proceedings 3rd International Symposium on Differential Satellite Navigation Systems, Canary Wharf, London, U.K., April 18– 22, 1994.

Stewart, M. P., M. Tsakiri, J. Wang, and J. F. Monico, IGEX—A regional analysis of data from the Southern Hemisphere,Proceedings of the IGEX-98 Workshop, Nashville, U.S.A., September 13–14, 1999.

Willis, P. and J. Slater, International GLONASS Experiment (IGEX-98), inIGS 1998 Annual Report, pp. 400–838, Jet Propulsion Laboratory, California Institute of Technology, Pasadena, California, JPL Publication, 1999.

Referências

Documentos relacionados

social assistance. The protection of jobs within some enterprises, cooperatives, forms of economical associations, constitute an efficient social policy, totally different from

Fractures were made in all samples. Figure 1 shows fractures of test bars cast from the examined high-aluminium iron in base condition and with the addition of vanadium

Na hepatite B, as enzimas hepáticas têm valores menores tanto para quem toma quanto para os que não tomam café comparados ao vírus C, porém os dados foram estatisticamente

É nesta mudança, abruptamente solicitada e muitas das vezes legislada, que nos vão impondo, neste contexto de sociedades sem emprego; a ordem para a flexibilização como

Representar este Oriente, (re)construí-lo discursivamente, era revivê-lo, era transportar para as terras de África uma glória que ficou na centúria de quinhentos. Muito

A sua presença no NARC, tal como em outros con‑ textos do casco urbano de Olisipo, onde se observa uma tendência para linearidade formal dos modelos cerâmicos de tradição romana

The probability of attending school four our group of interest in this region increased by 6.5 percentage points after the expansion of the Bolsa Família program in 2007 and

Essa questão é retomada por Arroyo (2004) quando reflete acerca das práticas educativas na relação do particular com o universal, destacando que defender o específico do